See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/315708427

Guidelines for Design and Operation of U-turns- Safety Evaluation of U-turn

Design

Technical Report · February 2017

CITATIONS READS

0 292

5 authors, including:

Maryam Shirinzad Raul Avelar

Texas A&M University Texas A&M University System
5 PUBLICATIONS   1 CITATION    34 PUBLICATIONS   72 CITATIONS   

SEE PROFILE SEE PROFILE

Bahar Dadashova
Texas A&M University
16 PUBLICATIONS   18 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

FURGOSEG Joint Project. P 24/08. Ministry of Public Works in its National Plan R&D 2008-2011 View project

Innovative Tools and Techniques in Identifying Highway Safety Improvement Projects: Technical Report View project

All content following this page was uploaded by Bahar Dadashova on 30 January 2019.

The user has requested enhancement of the downloaded file.

Article

Transportation Research Record

1–11
Ó National Academy of Sciences:
Safety Evaluation for Turnarounds at Transportation Research Board 2018
Article reuse guidelines:
Diamond Interchanges: Assessing the sagepub.com/journals-permissions
DOI: 10.1177/0361198118797186

Texas U-Turn journals.sagepub.com/home/trr

Karen K. Dixon1, Raul E. Avelar1, Maryam Shirinzadeh Dastgiri1,

and Bahar Dadashova1

Abstract
Texas frontage road turnarounds at diamond interchange locations are a common treatment in the State of Texas. This
configuration, also often referred to as the Texas U-turn, allows vehicles traveling on a one-way frontage road to access
the opposing direction one-way frontage road via a U-turn before the terminal intersections at the diamond interchange.
This technique removes two potential left-turn maneuvers from the adjacent at-grade intersections. The frontage road
turnaround has operational benefits that result from shifting vehicles that would typically be occupying the intersection.
The safety of these turnaround configurations, however, is not well known. This paper focuses on the observed safety
performance of these turnaround configurations at diamond freeway interchanges in Texas. The authors first developed a
large randomly sampled data set to facilitate a statistically reliable assessment of U-turn safety performance for Texas
interchanges. Next they conducted a safety performance assessment at locations with and without turnarounds and
determined that site features that significantly influence the number of crashes include the volume of cross street traffic,
the cross street right-turn configuration at frontage roads, the minimum radius in the turnaround, the longitudinal dis-
tance from the U-turn exit to the closest downstream driveway, and the number of lanes on each frontage road. The
safety performance findings at locations with and without turnarounds suggests that an agency could construct a turn-
around to enhance facility operations with the knowledge that construction will not adversely impact safety. Although
the research introduced in this paper focuses on Texas locations, the findings should be applicable to similar facilities in
other states if the turnaround conditions are similar.

Frontage road turnarounds are located at major
interchanges in Texas in urban and suburban regions (1).
For a turnaround to be a feasible option, the frontage
roads located parallel to the freeway lanes must have
one-way operations. At locations where the freeway
lanes are depressed, the turnaround is located on a
structure adjacent to the arterial cross street lanes. When
the freeway is elevated, the turnaround location will be
under a bridge as shown in Figure 1. For both
conditions, sight distance and visibility are important
considerations that influence the U-turn geometry.
In 2010, the American Association of State Highway
and Transportation Officials (AASHTO) published the
first edition of the Highway Safety Manual (2). This docu-
ment introduced the concept of integrating safety predic-
tion based on statistically reliable safety performance
functions that uniquely assess key characteristics of a road-
way countermeasure or configuration. Unfortunately,
safety research consistent with the techniques recom-
mended in the Highway Safety Manual is not yet available
for the Texas turnaround configuration. Consequently,
transportation professionals do not have the ability to
directly assess the number of expected crashes at an inter-
change location or determine how the configuration of the
U-turn can influence the number of crashes. The research
summarized in this paper, therefore, presents the results of
a recent Texas turnaround safety assessment performed on
behalf of the Texas Department of Transportation
(TxDOT). Although the analysis focuses on Texas inter-
changes, the findings from this study are likely to be
1
Texas A&M Transportation Institute, College Station, TX
Corresponding Author:
Address correspondence to Karen K. Dixon: k-dixon@tti.tamu.edu
2 Transportation Research Record 00(0)
Figure 1. Example of a Texas turnaround underpass configuration.
Source: Photo by Karen Dixon, Aerial View by GoogleEarth.
representative of similar interchange configurations in
other states.
Database Development
Currently, the State of Texas does not maintain a com-
prehensive intersection or interchange database. To over-
come this limitation, the authors first developed a study
sample by identifying candidate study sites using geo-
graphic information system (GIS) merging techniques
and then developing a geographically stratified random
sample for database development and subsequent safety
assessment.
Identifying Candidate Study Sites
The authors used existing roadway functional system
information identified in the TxDOT roadway inven-
tory database as an initial step toward determining free-
way and arterial networks where diamond interchanges
could be potentially located. Using GIS tools, the
authors identified locations where interstate, freeway,
or expressway facilities crossed arterial roadways. The
authors developed separate GIS shape files based on
these road categories and identified locations for poten-
tial points of the intersection that could represent inter-
changes or intersections. These potential intersections
included intersecting points as far as 200 ft (approxi-
mately 60 m) from the freeway centerline. Using manual
evaluations, the authors then confirmed which inter-
secting points did represent intersections between free-
ways and arterials. Figure 2 depicts these identified
intersections for the selected roadway network (freeway
and arterial).
Following this initial GIS screening for intersecting
points, the authors next matched the large number of
intersecting points (11,289 identified points) to potential
locations where an interchange was likely. To narrow the
search, an intersection with multiple points (generally
four to six points) located close together often repre-
sented some sort of interchange. As can be observed in
Figure 3a, an interchange with a U-turn will normally
have at least six intersecting points. Therefore, to identify
these potential interchange locations, the authors applied
an additional 300 ft (approximately 100 m) buffer to all
intersecting points as shown in Figure 3b to measure the
distance between two intersecting points of randomly
selected interchanges. As the objective of this activity
was to include interchanges with and without U-turns, a
diamond interchange with four or more intersection
points within this buffer region represented a potential
U-turn location. Finally, the authors filtered locations
and assigned a single consolidated intersecting point to
represent the latitude and longitude for the potential
study interchange.
Developing a Stratified Random Sample
The selection of interchange locations that represent the
larger diamond interchange population requires a ran-
dom selection of the study locations. As the State of
Texas is very large, the sampling procedure considered
geographic location so that the resulting sample could be
representative of the diamond interchanges across the
state. The final TxDOT project report provides
Dixon et al
Figure 2. Locating an interchange with a U-turn (a) aerial view and (b) applying 300 ft buffer.
considerable detail about this sampling technique, but
the final stratified random sample included 168 diamond
interchanges with U-turns and 60 prospective diamond
interchanges without U-turns. The authors used this data
set for the subsequent site selection and data collection
activities. The geographically stratified random sample
included a two-stage process in which the authors first
randomly selected interchanges from the entire state and
then randomly selected additional interchanges so that
the dataset included at least one interchange from every
TxDOT district with freeway facilities. The initial inter-
change identification utilized geographic shapefiles and
locations where arterials and freeway vectors crossed fol-
lowed by photographic confirmation of each diamond
interchange.
Data Collection and Database Development
Following the identification of the prospective study
sites, the authors acquired data for each candidate site
3
that included geometric data acquired from the TxDOT
roadway inventory database and was supplemented or
confirmed using aerials from Google Earth ProÒ. In
addition, the database included average annual daily
traffic (AADT), the K-factor (for converting daily traffic
volume proportions during peak hours), and the direc-
tional distribution factor known as the D-factor. This D-
factor proved to be a valuable way to confirm one-way
frontage road operations. In addition, the authors
assigned crash data acquired from the Texas Crash
Records Information System for the years 2009–2015.
The randomly selected study sites included a mixture
of locations with and without U-turns. For locations
with U-turns, the authors reviewed historic aerial photo-
graphs to confirm the presence of the U-turn for the
entire seven-year period. For locations where the U-turn
construction occurred during this study period, the
authors removed any unconfirmed crash years. As an
example, if an aerial photograph indicated a U-turn was
not present in 2010 but was present in 2013, the database
4 Transportation Research Record 00(0)
Figure 3. Locating an interchange with a U-turn (a) aerial view and (b) applying 300 ft buffer.
excluded crash data for 2011 and 2012. In addition,
depending on the date of the aerial photograph, 2010
and 2013 crash data may have also been removed if the
‘‘before’’ site aerial was not in December or the ‘‘after’’
site aerial was not in January of the study year.
The analysis database included many potential site
variables. Many of these did not prove to be significant
and are not included for the purposes of brevity (includ-
ing frontage road AADT, which the research team deter-
mined provided inaccurate values when compared to the
crossroad AADT). Key data elements influential to the
safety performance include the following data elements:
 Number of total lanes on each frontage road
approach;
 Longitudinal distance between the U-turn exit and
the first downstream driveway;
 Presence of traffic signal at the study intersections.
 Arterial right-turn treatment onto frontage roads.
 Depressed or elevated U-turn configuration; and
 U-turn leg dimensions (radii, lane widths, and
length of dedicated U-turn).
In some cases, the data reduction process could not
determine all site characteristics for a location. Often this
constraint was a result of the interchange configuration.
For example, U-turn leg dimensions could not always be
determined at underpass turnaround locations because of
occlusion from the structure. The authors elected to retain
these sites and use a varying sample size during the statis-
tical analysis stepwise variable assessment. The following
sections review each of these roadway characteristics.
Number of Total Lanes on Each Frontage Road Approach. For
each side of a study site (referred to from this point
forward as a half site), the authors determined the num-
ber of frontage road main lanes at the intersection
approach. Figure 4 depicts the total number of lanes as
FRA and FRB. The authors used the labels ‘‘A’’ and ‘‘B’’
to identify each half site for documentation purposes.
They do not imply specific directional information. It
should be noted that most frontage roads had two to
three lanes near the U-turn locations. It should also be
noted that the study included 203 half sites with turn-
arounds and 133 half sites that did not have turnarounds.
This resulted in a total of 336 half sites.
Longitudinal Distance Between the U-turn Exit and the First
Downstream Driveway. The longitudinal distance to the first
downstream driveway, measured from the point on the
frontage road where the U-turning vehicles exit the U-turn
and merge onto the frontage road, can influence the weav-
ing and merging conditions on the frontage road. This
label is represented as DWYA and DWYB in Figure 4. For
this study, the authors measured this distance for the drive-
ways that were within 500 ft of the U-turn exit. Out of 203
half sites, 53 of the locations had driveways positioned out-
side of this 500 ft threshold. The authors assigned a default
DWY value of 500 ft for these locations.
Presence of a Traffic Signal at Study Intersection. The authors
used aerial photographs and the companion StreetView
tool to determine the type of traffic control present at
each frontage road and cross street intersection location.
At locations with a traffic signal, this method did not per-
mit acquiring signal timing information. Approximately
93% of the intersections had traffic signals. Ultimately,
the authors included only signalized intersection loca-
tions in the subsequent safety assessment.
Dixon et al
Figure 4. Turnaround configuration and influential site characteristics.
Arterial Right-Turn Treatment onto Frontage Roads. As a vehi-
cle that is turning right from the cross street onto the one-
way frontage road may potentially encounter a conflict
with a vehicle exiting the U-turn, the authors categorized
the configuration of the cross street right-turn treatment
‘‘zone’’ for each one-way frontage road location. Each
configuration included the following two variables:
 Right-turn treatment ‘‘zone’’ entrance (see
Figure 5), and
 Right-turn treatment ‘‘zone’’ exit (see Figure 6).
The entrance treatments included shared and exclusive
cross street right-turn lanes with and without channeliza-
tion islands. Ultimately, the authors tested the influence
of these entrance treatments and determined that only
Option A (a shared right with no island) was significantly
associated with fewer crashes.
The exit treatment configurations required vehicles to
merge into existing travel lanes or the geometric config-
uration included an additional lane (either as an accelera-
tion lane or as another full lane). Even though most of
the study locations had traffic signals, exit treatments at
locations with islands often included supplemental traffic
control (usually a yield or stop sign).
Depressed or Elevated Interchange Configuration. Of the 336
half site locations, 300 (150 of 168 sites) of them had
5
arterial crossroads located below a freeway bridge, and
36 (18 of 168 sites) were elevated above the freeway.
Although the placement as elevated versus depressed did
not prove to be significant to intersection safety, the
roadway geometry was directly influenced by these
factors.
U-Turn Leg Dimensions (Widths and Lengths). Owing to the
orientation of the U-turns, interchange bridges, and simi-
lar characteristics, the authors acquired, where feasible,
the turning radius for the U-turn, the turning bay length,
and the lane length. Although the authors evaluated the
significance of all of these candidate variables, the mini-
mum U-turn radius proved to be the only critical variable
that influenced safety performance.
Analysis and Results
The authors performed a statistical analysis for the total
crash condition (at turnaround locations) as well as the
injury crash condition (denoted as the KAB model in
which ‘‘K’’ represents a fatal crash, ‘‘A’’ represents an
incapacitating injury crash, and ‘‘B’’ represents a non-
incapacitating injury crash). The analytical approach
used for this analysis was consistent with that used in a
variety of similar studies (3–5). The following section first
reviews the descriptive statistics for the larger dataset,
6 Transportation Research Record 00(0)

Table 1. Summary of Data Characteristics for a Site-Period

Variable name Mean SD Min Max Total N

Summary of observed crashes

Total crashes 25.59 23.53 0 118 25,999 1016
Total crashes (known coordinates) 25.38 23.36 0 117 25,783 1016
KAB crashes 7.22 6.87 0 41 7,336 1016
Crashes involving a left-turning vehicle from frontage road 3.72 5.24 0 34 3,777 1016
Crashes involving a left-turning vehicle from frontage road (known coordinates) 3.69 5.23 0 34 3,754 1016
KAB crashes involving a left-turning vehicle from frontage road 0.89 1.47 0 12 908 1016
KAB crashes involving a left-turning vehicle from frontage road (known coordinates) 0.89 1.46 0 11 903 1016
Additional variable characteristics
Minimum posted speed limit 45.4 5.9 30 55 46,145 1016
Maximum posted speed limit 46 6 30 55 46,705 1016

Note: Min = minimum; Max = maximum; SD = standard deviation; N = number.

Figure 6. Right-turn treatment ‘‘zone’’ exit options (a) add lane,

no additional control; (b) merge, yield control; (c) merge, stop
control; (d) merge, signal control; (e) merge, no additional
control; and (f) add lane, yield control.

Figure 5. Right-turn treatment ‘‘zone’’ entrance options
(a) shared right, no island, (b) exclusive right, no island, (c)
exclusive right, painted island, (d) exclusive right, raised island, and
(e) shared right, raised island, large radius.
then identifies some of the statistical modelling challenges
addressed, and finally presents the resulting statistical
models. Both models predicted crashes for the non-
freeway crashes at the entire diamond interchange (rather
than crashes on just one side).
Descriptive Statistics
Based on the premise that the number of left-turn crashes
that originated on a frontage road can be expected to be
reduced at U-turn locations, the authors inspected the
crash data (for the years 2009–2015) and site data using a
Dixon et al
variety of target crash conditions including the propor-
tion of crashes in which the vehicle originated on the
frontage road. In some cases, missing data required slight
modifications in sample sizes. This occurred when a data
element could not be determined using the aerial photos
and the roadway inventory file. For the initial inspection,
the data included 2019 site-periods (seven potential years
per site) from 164 sites; however, 77 of the site-periods
did not have available speed limit data and so the authors
could not evaluate these sites relative to speed limit. This
resulted in 1016 site-periods from 152 sites with turn-
arounds (see Table 1).
Statistical Analysis
Following the initial data inspection, the authors per-
formed a stepwise regression analysis to assess the influ-
ence of significant variables on the total number of
crashes, as well as crash severity for the dedicated U-turn
lane locations. During this process, the authors noted
that only a small number of the interchanges had a
posted speed limit of 30 mph. Most of the intersections
had speed limits of 35 to 55 mph with an average posted
speed limit of 45 to 46 mph. None of the 30 mph intersec-
tions had traffic signals for both of the intersections
associated with the interchange. To mitigate the influ-
ence of this subset of study sites, the authors elected to
remove these intersections from the subsequent analysis.
Model Development. The crash model development
focused on the identification of statistically significant
variables and an assessment of the best functional form
that represents each identified variable. For example, the
use of logarithmic adjustments may be appropriate for
some variables that do not follow a linear format when
graphically plotted. The authors used generalized mixed-
effects models to compensate for more than one study
year of data at a specific site location. The authors also
explored an extensive number of variables during the
variable selection. Of particular note is the selected traf-
fic volume variable. It is feasible to consider freeway
NTotal = e½5:304ð0:271 3 RtAÞ + ð0:703 3 AvgLnÞð0:268 3
volume, cross street volume, frontage road volume (for
both frontage roads), and U-turn volume. Ultimately,
the cross street AADT values proved to be the most
influential traffic volume variable for this assessment.
e½5:491ð0:271 3 RtAÞ + ð0:703 3 AvgLnÞð0:0017 3 DWY Þ + ð1:124 3 10 Þ 3 CSAADT 
NTotal =
7
The authors used a negative binomial model approach
for this analysis.
Total Crash Model. The authors used a ‘‘sample expansion’’
step before switching subsets of potential independent vari-
ables as the initial model development approach (6–8). For
the development of the final predictive model, the authors
focused on signalized intersection locations with speed lim-
its greater than 30 mph. Through the application of step-
wise regression procedures, the authors then developed a
final predictive model for total crashes. The model selec-
tion occurred in several stages so that the authors could
consider groups of variables jointly available for data sub-
sets. Once a modelling stage had arrived at the point with
the best fit and the fewest variables (referred to as a parsi-
monious model), the authors fit the model to the largest
subset of data that included all the variables in the model
specification. This iterative process continued until the
authors had considered all variables at least two times for
inclusion into the model. The resulting dataset represented
86 site locations with 459 site-periods available for model
estimation. Table 2 shows the final total crash model. The
authors applied scaling adjustments to two of the variables
with more widely dispersed data (9–11). The cross street
right-turn entrance treatment A (shared right with no
island) proved to be more significant that the other
entrance treatments. The cross street right-turn exit treat-
ments were not statistically significant for the total crash
model.
A common goodness of fit assessment gives the cumu-
lative residual (CURE) plot. Optimally, the CURE plot
for each variable should oscillate around the line that
represents zero. For the total crash model, these plots
depict minimal deviations beyond the expected bound-
aries for key variables (see Figure 7). As can be observed,
the expected oscillation occurs for the lower volume
thresholds, but the cumulative residuals appear to be
skewed for the higher volume locations.
The final model presents a functional form that incor-
porates scaling of some variables. This total crash model
is
13, 516:82
Þ + 0:113 3 ½CSAADT
DWY196:319
155:752 10, 059:57 (0:951 3 ln(Rmin))
The equation included a scaling adjustment for the
DWY and the CSAADT variables. The authors adjusted
these scaled variables by subtracting the mean value and
then dividing by the standard deviation of the variable.
The reduced NTotal model is:
5
Rmin0:9512
8 Transportation Research Record 00(0)

Table 2. Predictive Model for Total Crashes (Signalized Sites)

Continuous variable descriptive statistics

Variable name Description Mean SD Min. Max. Total N

AvgLn Average number of 2.6 0.5 2 4 1195 459

frontage
road lanes FRA and FRB
(see Figure 4)
DWY Distance to closest 196.3 155.8 10 500 90,110 459
downstream driveway
(ft)
CS_AADT Cross street AADT (vpd) 13,516.8 10,059.6 200 54,609 6,204,220 459
Rmin Minimum turning radius in 49.9 14.6 22 129 22,918 459
U-turn (ft)

Total crash model

Standard
Variables Estimate error Z value Pr (.|z|) Significanceb

(Intercept)a 5.3041 1.0862 4.8834 1.0428 x 1026 ***

RtA 20.2708 0.1023 22.6480 0.0081 **
AvgLn 0.7027 0.1616 4.3490 0.0000 ***
Scale 20.2684 0.0719 23.7320 0.0002 ***
(D_to_closest_driveway)
Scale (CS_AADT) 0.1131 0.0489 2.3120 0.0208 *
ln (Rmin) 20.9512 0.2454 23.8760 0.0001 ***

Note: Min = minimum; Max = maximum; SD = standard deviation; N = number; RtA = Number of instances at the site where RtA had a shared right turn
lane and no channelization island (see Figure 5). Value of RtA ranges from zero (no shared lane option) to two (shared lane option at both cross street
right-turn locations).
a
Includes adjustment random effects.
b
Significance levels are as follows:
*
Statistically different from 0.0 at the 5.0% significance level.
**
Statistically different from 0.0 at the 1.0% significance level.
***
Statistically different from 0.0 at the 0.1% significance level.

One interesting observation about the final (and ulti-
mately the KAB) model is that the presence of a turn-
around does not appear as a critical variable in the
model. The authors included this variable in the stepwise
analysis and it was not significant. This finding suggests
that constructing a turnaround does not significantly
affect the total number of non-freeway interchange
crashes, and so this treatment should complement opera-
tional benefits of adding turnarounds at diamond inter-
change locations.
By inspection of the variables included in the total
crash model for interchanges with turnarounds, the fol-
lowing general observations merit consideration:
 Locations where the right-turn from the cross
street originates from a shared lane and does not
have a large turning radius or a raised island
(Option A) could result in 23.7% fewer crashes
(calculated as 12e(20.2708) = 0.237). This finding
is significant at 1.0%.
 For frontage roads with 2–4 lanes, the number of
crashes increases by a factor of 2.01 (doubles) for
each additional frontage road lane (calculated as
e(0.7027) = 2.014). This finding is significant at
0.1%. This finding is likely a surrogate for the
varying frontage road AADT values.
 The number of crashes reduces as the distance to
the closest downstream driveway increases. This
reduction is approximately 1.7% for each addi-
tional 10 feet between the closest U-turn exit and
the downstream driveway (calculated as
0:2684
1  e 155:752 3 10 = 0:0171). This finding is significant
at 0.1%.
 The number of crashes increases by 1.1% for each
additional 1000 vehicles per day (vpd) increase in
cross street AADT (calculated as
0:1131
1  e10, 059:6 3 1000 = 0:011). This finding is significant
at 5.0%.
 The number of crashes decreases by 8.7% for each
increase in 10% in the turning radius of the
Dixon et al 9

200

200
Number of Crashes
Number of Crashes

100

100
0

0
-100

-100
-200

-200
0 10000 20000 30000 40000 50000 1.0 1.5 2.0 2.5 3.0

Cross Road AADT (vpd) Min No Lanes Frontage

200

200
Number of Crashes

Number of Crashes
100

100
0

0
-100

-100
-200

-200

0 100 200 300 400 500 20 40 60 80 100 120

D_to_Closest_Driveway MinR

Figure 7. CURE Plots for the total crash model.

U-Turn (calculated as 1  e0:9512 3 lnð1:1Þ =  Sites where the right-turn from the cross street
0:0867). This finding is significant at 0.1%. must share a lane have 26.4% fewer KAB crashes
(calculated as 12e(20.3063)). This finding is signifi-
cant at 1.0%.
KAB Crash Model. The authors used a similar modelling  Sites where the right-turn traffic must merge with
approach as that applied for total crashes to evaluate the the frontage road traffic (without adding a lane)
injury crashes at the Texas turnaround locations (see have fewer severe crashes (significant at \
Table 3). The CURE plots for this model exhibited simi- 0.01%). This merge configuration is associated
lar oscillating patterns to those as shown for total crashes with a reduction of 16.4% (calculated as
and therefore are not shown for this model. 12e(20.1801)). This finding is significant at 1.0%.
The KAB crash model equation is:

NKAB = e½3:08ð0:31 3 RtAÞ + ð0:48 3 AvgLnÞð0:18 3 MergeRT Þð0:20 3 Þ + ð0:072 3 lnðCS

DWY 191:58
151:86 AADT Þ)(0:67 3 ln(Rmin))

After simplification, the resulting predictive KAB  The number of KAB crashes is smaller by 1.3%
equation takes the following form: for each additional 10 feet between the closest

CS AADT 0:0719
NKAB = 3 e½3:32ð0:31 3 RtAÞ + ð0:48 3 AvgLnÞð0:18 3 MergeRT Þð0:0013 3 DWY Þ
Rmin0:6684

By inspection of the variables included in the KAB downstream driveway and the U-turn exit (calcu-
model for signalized interchanges with turnarounds, the 0:1969
lated as 1  e 151:86 3 10 = 0:0126). This finding is
following observations merit consideration: significant at 1.0%.
10 Transportation Research Record 00(0)

Table 3. Predictive Models for KAB Crashes (Signalized Intersections)

Continuous variable descriptive statistics for the KAB model

Variable name Description Mean SD Min. Max. Total N

CS_AADT Cross street AADT (vpd) 13,872.4 10,442.3 200 54,309 6,478,395 467
AvgLn Average number of frontage 2.62 0.45 2 4 1225 467
road lanes FRA and FRB (see
Figure 4)
DWY Distance to closest downstream 191.54 151.86 10 500 89,467 467
driveway (ft)
Rmin Minimum turning radius in U- 49.99 14.84 22 129 23,344 467
turn (ft)

Final KAB model

Variables Estimate Standard error Z value Pr (.|z|) Significanceb

(Intercept)a 3.0758 1.1096 2.7720 5.5718E-03 .

ln (CS_AADT) 0.0719 0.0392 1.8360 6.6290E-02 .
RtA 20.3063 0.1013 23.0220 2.5100E-03 **
AvgLn 0.4797 0.1580 3.0350 2.4000E-03 **
MergeRT 20.1801 0.0892 22.0200 4.3350E-02 *
Scale (DWY) 20.1969 0.0699 22.8190 4.8200E-03 **
ln (Rmin) 20.6684 0.2412 22.7710 5.6000E-03 **

Note: Min = minimum; Max = maximum; SD = standard deviation; N = number; RtA = Number of instances at the site where right-turn ‘‘zone’’ entrance
treatment had a shared right turn lane and no channelization island (see Figure 5). Value of RtA ranges from zero (no shared lane option) to two (shared
lane option at both cross street right-turn locations). MergeRT = Number of instances at the site where the right-turn ‘‘zone’’ exit treatment merged into
an existing lane (see Figure 6). Value of MergeRT ranges from zero (only included added lanes) to two (all cross street right-turn lanes require vehicles to
merge into an existing lane).
a
Includes adjustment random effects.
b
Significance levels are as follows:
*
Statistically different from 0.0 at the 5.0% significance level.
**
Statistically different from 0.0 at the 1.0% significance level.
***
Statistically different from 0.0 at the 0.1% significance level.

Table 4. Influence of Site or Traffic Characteristics on Crashes

Significant influence
Site or traffic characteristic Total crashes KAB crashes

As the cross street AADT increases: Total crashes increase KAB crashes increase
Cross street right-turn maneuvers onto the Fewer total crashes Fewer KAB crashes
frontage road that originate in a shared lane
are associate with:
Cross street right-turning vehicles that merge No significantly noticeable change in total crashes Fewer KAB crashes
into existing lanes are associated with:
As the number of lanes increases for each Total crashes increase KAB crashes increase
frontage road approach:
As the longitudinal distance between the U-turn Total crashes decrease KAB crashes decrease
exit to the nearest downstream driveway
increases:
As the minimum U-turn radius increases: Total crashes decrease KAB crashes decrease

 The number of KAB crashes is smaller by 6.2%

Conclusions and Recommendations
for each increase of 10% in the turning radius of This study examined the safety effects of turnarounds at
the U-Turn (calculated as 1  e0:6684 3 lnð1:1Þ diamond interchanges in Texas. Based on a statistical eva-
= 0:62). This finding is significant at 1.0%. luation of total crashes and injury crashes, the authors
 Dixon et al
generally concluded that key variables associated with
crash frequency include cross street AADT, cross street
right turn configuration, number of frontage road lanes,
longitudinal distance from U-turn exit to nearest down-
stream driveway, and U-turn minimum radius values col-
lectively influence the number of crashes at turnaround
locations. The authors limited the study sites to locations
with posted speed limits on the frontage road of 35 to
55 mph and with signalized intersection configurations.
Table 4 summarizes how each of these site or traffic char-
acteristics influences the number of predicted crashes.
In addition, the authors evaluated sites with and with-
out turnarounds. During the statistical analysis, it
became clear that the variable that indicated the presence
of a turnaround was not statistically significant. This
finding suggests that the construction of turnarounds at
diamond interchanges will not substantially affect the
total number of crashes at these locations. This means
that the construction of a turnaround as a mechanism for
improving operations and removing the two left-turns
from adjacent signalized intersections will not have any
significant adverse safety implications. In Texas, the con-
struction of turnarounds is routine for newly constructed
or retrofitted interchanges in urban areas. The findings
of this study demonstrate that the construction of turn-
arounds for operational purposes will not introduce addi-
tional safety concerns.
Acknowledgments
The authors would like to thank the Texas Department of
Transportation for their support of this research effort.
Author Contributions
The authors confirm contribution to the paper as follows: study
conception and design: K. Dixon, R. Avelar; data collection: K.
Dixon, M. Shirinzadeh Dastgiri, B. Dadashova; analysis and
interpretation of results: K. Dixon, R. Avelar, M. Shirinzadeh
Dastgiri, B. Dadashova; draft manuscript preparation: K.
Dixon, R. Avelar. All authors reviewed the results and
approved the final version of the manuscript.
View publication stats
11
References
1. TxDOT. Roadway Design Manual, Revised October 2014.
http://onlinemanuals.txdot.gov/txdotmanuals/rdw/
rdw.pdf. Accessed February 20, 2017.
2. American Association of State Highway Transportation
Officials. Highway Safety Manual. American Association
of State Highway Transportation Officials, Washington,
D.C., 2010.
3. Bauer, K. M., D. W. Harwood, and K. R. Richard. Safety
Effects of Using Narrow Lanes and Shoulder-Use Lanes to
Increase the Capacity of Urban Freeways. Report No.
FHWA-HRT-05-001. Federal Highway Administration,
McLean, VA, 2005.
4. Bauer, K. M., D. W. Harwood, W. E. Hughes, and K. R.
Richard. 2004. Safety Effects of Using Narrow Lanes and
Shoulder-Use Lanes to Increase the Capacity of Urban
Freeways. Transportation Research Record: Journal of the
Transportation Research Board, 2004. 1897: 71–80.
5. Bonneson, J. A., S. Geedipally, M. Pratt, and D. Lord.
NCHRP Project No. 17–45: Safety Prediction Methodology
and Analysis Tool for Freeways and Interchanges. Trans-
portation Research Board of the National Academies,
Washington, D.C., 2012.
6. Fox, J., and S. Weisberg. An {R} Companion to Applied
Regression, 2nd ed. SAGE Publications, Thousand Oaks,
CA, 2011.
7. Hauer, E., and J. Bamfo. Two Tools for Finding what
Function Links the Dependent Variable to the Explana-
tory Variables. Proc., ICTCT 1997 Conference. Lund, Swe-
den, 1997.
8. Hilbe, J. M. Negative Binomial Regression, 2nd ed. Cam-
bridge University Press, Cambridge, United Kingdom,
2011.
9. Park, E., J. Park, and T. Lomax. A Fully Bayesian Multi-
variate Approach to Before-After Safety Evaluation. Acci-
dent Analysis and Prevention, Vol. 42, No. 4, 2010, pp.
1118–1127.
10. Persaud, B., B. Lan, C. Lyon, and R. Bhim. Comparison
of Empirical Bayes and Full Bayes Approaches for Before-
After Road Safety Evaluations. Accident Analysis and Pre-
vention, Vol. 42, No. 1, 2010, pp. 38–43.
11. Ramsey, F. L., and D. W. Schafer. The Statistical Sleuth:
A Course in Methods of Data Analysis, 2nd ed. Duxbury,
Pacific Grove, CA, 2002.
The Standing Committee on Operational Effects of Geometrics
(AHB65) peer-reviewed this paper (18-04662).